Holoceno Medio Corales Australia

8/12/2019 Holoceno Medio Corales Australia

1/6

DOI: 10.1126/science.279.5353.1014, 1014 (1998);279Science

, et al.Michael K. GaganTropical Western PacificTemperature and Surface-Ocean Water Balance of the Mid-Holocene

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by, yo u can order high-quality copies for yourIf you wish to distribute this articl e to others

here.following the guidelines

can be obtained byPermission to republis h or r epurpose articles or portions of articles

): September 21, 2011www.sciencemag.org (this infomation is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/279/5353/1014.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/279/5353/1014.full.html#ref-list-1, 8 of which can be accessed free:cites 25 articlesThis article

188 article(s) on the ISI Web of Sciencecited byThis article has been

http://www.sciencemag.org/content/279/5353/1014.full.html#related-urls20 articles hosted by HighWire Press; see :cited byThis article has been

http://www.sciencemag.org/cgi/collection/atmosAtmospheric Science

subject collections:This article appears in the following

registered trademark of AAAS.is aScience 1998 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005.(print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience
http://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/content/279/5353/1014.full.htmlhttp://www.sciencemag.org/content/279/5353/1014.full.htmlhttp://www.sciencemag.org/content/279/5353/1014.full.htmlhttp://www.sciencemag.org/content/279/5353/1014.full.html#ref-list-1http://www.sciencemag.org/content/279/5353/1014.full.html#ref-list-1http://www.sciencemag.org/content/279/5353/1014.full.html#ref-list-1http://www.sciencemag.org/content/279/5353/1014.full.html#ref-list-1http://www.sciencemag.org/content/279/5353/1014.full.html#related-urlshttp://www.sciencemag.org/content/279/5353/1014.full.html#related-urlshttp://www.sciencemag.org/content/279/5353/1014.full.html#related-urlshttp://www.sciencemag.org/content/279/5353/1014.full.html#related-urlshttp://www.sciencemag.org/cgi/collection/atmoshttp://www.sciencemag.org/cgi/collection/atmoshttp://www.sciencemag.org/cgi/collection/atmoshttp://www.sciencemag.org/content/279/5353/1014.full.html#related-urlshttp://www.sciencemag.org/content/279/5353/1014.full.html#ref-list-1http://www.sciencemag.org/content/279/5353/1014.full.htmlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtl


2/6

J. Gorosabel et al. , in preparation.21. P. Me sza ros and M. J. Rees, Astrophys. J. 405 , 278

(1993); ibid. 418 , L59 (1993).22. , ibid. 476 , 232 (1997).23. The impulsive fireball model in ( 22 ) predicts that the

observed break frequency m crosses the opticalband at a time t op 3 10 3 t , where t is theduration of thegamma rayflash. Takingintoaccountthat the bulk of the GRB 970508 gamma ray emis-sion lasted 3.6 s, t op 3 hours, which is muchshorter than the observed t op 2 days.

24. R. A. M. J. Wijers, M. J. Rees, P. Me sza ros, Mon.Not. R. Astron. Soc. 288 , L51 (1997)

25. J.I. KatzandT. Piran, Astrophys. J. 490 , 772 (1997)and references therein.

26. Like GROJ0422 32[A. J.Castro-Tirado,J. L.Ortiz,J. Gallego, Astron. Astrophys. 322 , 507 (1997)] andGRO J1655-40 [J. A. Orosz, R. A. Remillard, Ch. D.Bailyn, J. E. McClintock, Astrophys. J. 478 , L83(1997)].

27. S. K. Chakrabarti and P. J. Wiita, Astrophys. J. 411 ,

602 (1993). See also J. J. Brainerd, ibid. 394 , L33(1992).

28. M. Vietri, ibid. 488 , L105 (1997).29. The fastest x-ray flare ever seen in an extragalactic

object was observed in May 1994 in the BL Lacobject PKS 2155-304. The x-ray flare was followedby a broader, lower amplitude extreme-UV flare 1day later and a broad, low-amplitude UV flare 2days later [E. Pian et al. , Astrophys. J. 486 , 784(1997)].

30. M. Tavani, Astrophys. J. 466 , 768 (1996); E. Wax-man, ibid. 485 , L5 (1997).

31. T. Galama et al. , IAU Circ. 6655 (1997).32. M. R. Metzger et al. , IAU Circ. 6676 (1997).33. S. Djorgovski et al. , IAU Circ. 6658 (1997); IAU Circ.

6660 (1997); B. Schaefer et al. , IAU Circ. 6658(1997); P. Groot et al. , IAU Circ. 6660 (1997); C.Chevalier and S. Ilovaisky, IAU Circ. 6663 (1997); A.Kopylov etal. , ibid.; IAUCirc.6671 (1997); A.Fruchterand L. Bergeron, IAU Circ. 6674 (1997); (31, 32 ).

34. We are grateful to M. Santos-Lleo , M. Mas-Hesse, B.Montesinos, P. Rodr guez-Pascual, L. Sa nz Fernan-

dez de Co rdoba, E. Fernandes-Vieira, J. R. Acarreta, A. Alberdi, M. Cervino, T. Galama, N. Schartel, D. Thompson, and R. Wijers for fruitful conversations,andto J. Gallego, J. Zamorano, andM. Ferna ndezfortheir good will. We are grateful to the director of NOT for his continued support of this program, and to J.Perez-Mercador for his encouragement. J. Greiner issupported by the Deutsche Agenturfu r Raumfahrtan-gelegenheiten. C.B., A.G., N.M., and A.P. acknowl-edge the support of the Universita di Bologna. TheBeppoSAX team acknowledges the Italian Space Agency for support. The Calar Alto German-SpanishObservatory is operated jointly by the Max-Planck In-stitut fu r Astronomie in Heidelberg and the Comisio nNacional de Astronom a, Madrid. The WHT is operat-ed by the Royal Greenwich Observatory in the Span-ish Observatorio del Roque de los Muchachos of theInstituto de Astrof sica de Canarias. The Loiano Tele-scope is operated by Osservatorio Astronomico diBologna and Universita di Bologna.

24 June 1997; accepted 14 January 1998

Temperature and Surface-Ocean Water Balanceof the Mid-Holocene Tropical Western Pacific

Michael K. Gagan, Linda K. Ayliffe, David Hopley,Joseph A. Cali, Graham E. Mortimer, John Chappell,Malcolm T. McCulloch, M. John Head

Skeletal Sr/Ca and 18 O/ 16 O ratios in corals from theGreat Barrier Reef, Australia, indicatethat the tropical ocean surface 5350 years ago was 1C warmer and enriched in 18 Oby 0.5 per mil relative to modern seawater. The results suggest that the temperatureincrease enhanced the evaporative enrichment of 18 O in seawater. Transport of part ofthe additional atmospheric water vapor to extratropical latitudes may have sustained the18 O/ 16 O anomaly. The reduced glacial-Holocene shift in seawater 18 O/ 16 O ratio pro-duced by the mid-Holocene 18 O enrichment may help to reconcile the different tem-perature histories for the last deglaciation given by coral Sr/Ca thermometry and fora-miniferal oxygen-isotope records.

O xygen isotopes in foraminifera fromdeep-sea cores have been used successfullyto reconstruct continental ice volumes, gla -cio-eustatic sea level, and deep -ocean tem -peratures throughout the last glacial -inter -glacial cycle (1). Yet there is still consider -able debate regarding the relative contribu -tions of changes in the oxygen isotopiccomposition of seawater and ocean temper -ature to the change in foraminiferal 18 Ovalues (2) in the tropical surface ocean. Theamplitude of the glacial -Holocene 18 O

shift for planktonic foraminifera in thetropics is about the same as that for forami -nifera living in the near -constant tempera -ture environment of the deep ocean ( 3).The apparent similarity between the tworecords has led to the interpretation that

the tropical surface ocean cooled by nomore than 2C during the last glacial max -imum (LGM), an interpretation rein -forced by the conclusions of the CLIMAPproject ( 4). Recent measurements of Sr/Ca ratios in corals from Barbados ( 5) andVanuatu ( 6) indicate that both the tropi -cal Atlantic and Pacific oceans cooled byup to 5C during the LGM. Despite theconsistency of this cooling with the 5to8C cooling estimated for the tropical at -mosphere ( 7), the coral records have not

been widely accepted because the glacial-

Holocene 18 O shift recorded by plank -tonic foraminifera cannot accommodatean ocean temperature shift of 5C, partic -ularly in the tropical Pacific where the

18 O shift is anomalously small ( 8). If the coral Sr/Ca thermometer is reliable,the relatively small 18 O shift for Pacificplanktonics may be a result of undetectedchanges in the distribution of 18 O in thesurface ocean ( 9). Therefore, obtainingaccurate estimates of the past 18 O distri -bution in tropical surface waters, by some

independent means, is crucial to establish -ing the role of the tropical oceans in glob -al climate change.

In principle, precise measurements of Sr/

Ca ratios and 18

O values in coralline ara-

gonite should make it possible to determineuniquely the past oxygen isotopic compo -sition of seawater, by removal of the tem -perature component of the coral 18 O sig-nal. However, the temperature depen -dence of coral Sr/Ca ratios has been ques -tioned recently ( 10) on the basis thatbiological controls on Sr and Ca uptakecan induce uncertainties of up to 3C inreconstructed temperatures. Here, we firstverify the reliability of the coral Sr/Cathermometer for several modern coralsgrowing in suboptimal environmental set -tings like those that may prevail duringglacial-interglacial transitions when coralscolonize transient shorelines marked byfluctuations in temperature, salinity, andwater turbidity. Next we demonstrate astrong correlation between coral Sr/Caand 18 O and, by coupling these measure -ments, determine the surface temperatureand 18 O of western Pacific seawater soonafter the end of the last deglaciation. In -terpretations of the glacial -Holocene shiftin foraminiferal 18 O have been madewith the assumption that the 18 O distri -bution in the surface ocean was the sameas today at 6000 years before the

present, by which time the discharge intothe ocean of 18 O -depleted polar meltwaterwas essentially complete ( 11). We test thisassumption by analyzing a 5350 -year B.P.coral from the Great Barrier Reef.

We measured skeletal Sr/Ca ( 12) forcolonies of Porites lutea growing in threedifferent oceanic environments includingthe Great Barrier Reef, the eastern IndianOcean, and the Indonesian seaway. Thesetest sites encompass seasonal sea surfacetemperature (SST) ranges (20 to 31C)spanning much of the survival range for

M.K. Gagan,L. K.Ayliffe, J.A. Cali, G.E. Mortimer, M.T.McCulloch, Research School of Earth Sciences, Austra-lian National University, Canberra, ACT 0200, Australia.D. Hopley, Sir George Fisher Centre, James Cook Uni-versity, Townsville, Queensland 4811, Australia.J. Chappell and M. J. Head, Research School of Pacificand Asian Studies, Australian National University, Can-berra, ACT 0200, Australia.

SCIENCE VOL. 279 13 FEBRUARY 1998 www.sciencemag.org1014


3/6

Porites and cover the entire range of SSTsdefined by fossil corals for the last degla-ciation ( 5, 6). Moreover, the corals areexposed to seasonal extremes in salinity(28 to 39 per mil), coastal upwelling, andchanges in water column turbidity, all of which might affect the uptake of Sr andCa by coral skeletons ( 13) (Table 1). Wetook several precautions to avoid system -atic errors in comparing skeletal Sr/Cadata among corals and in comparing coraldata with instrumental data ( 14). Wematched the coral Sr/Ca ratios and instru -mental SST records ( 15) by first convert -ing the coral data to an accurate time

series, as described (16) (Fig. 1). We thenperformed least -squares regressions of cor-al Sr/Ca ratios and instrumental SSTsusing the data points defining the temper -ature maxima and minima for each coralrecord, totaling 16 years. The average dif -ference between the three Sr/Ca -SSTregression equations and their mean is0.3C over the entire SST range of 20 to31C (Fig. 2). As a result, the mean of thethree regression equations justifiably pro -vides a single Sr/Ca-SST relation for P.lutea defined by T 168.2 [15,674 (Sr/Ca) atomic ]. This relation is essentially thesame as the original equations derived by

Smith et al. (17) and Beck et al. (12) and,above 26C, the relation matches morerecent calibrations for P. lutea and P. lo-bata (18). Differences of up to 3C be -tween these relations and others derivedfor the genus Porites (13, 19) remain to beresolved.

To assess the effect of calcification rateon skeletal Sr/Ca ratios and 18 O values(20), we examined two P. lutea specimensgrowing at different rates in the same reef environment at Orpheus Island, centralGreat Barrier Reef. The mean annual ex -tension rates (12 and 22 mm year 1 ) andcalcification rates (1.3 and 2.5 g cm 2

year 1 ) (21) span a large portion of thespectrum of growth rates reported for Porites(22). The degree of correlation between theTable 1. Summary of study site locations, range of coral growth rates, and attributes of suboptimal

growth environments for P. lutea .

Location Extension(mm year 1 )Density

(g cm 3 )

Calcification(g cm 2year 1 )

SST (C)

Salinity(per mil) Environment

Nusa Barung Island,Java

24 1.17 2.8 20 to 30 30 to 34 Monsoonalrainfall, coastalupwelling

Orpheus Island, centralGBR

22 1.15 2.5 22 to 29 28 to 36 Monsoonal riverrunoff 12.5 1.08 1.3

Dampier Archipelago,eastern Indian Ocean

8 1.13 0.9 20 to 31 35 to 39 Seasonalhypersalinity

Fig. 1. ( A to C ) Comparison be-tween coral Sr/Ca in P. lutea (E )and blended ship- and satellite-de-rived SSTs (solid curves) ( 15 ) forNusa Barung Island, southeastJava (831 S, 11322 E); OrpheusIsland, central Great Barrier Reef

(1845 S, 14629 E); and Dampier Archipelago, eastern Indian Ocean(2036 S, 11645 E). The Dampiercoral grew in shallow nearshorewa-ters where the seasonal range inSST is larger than that recorded bythe1 latitude-longitudegridsof thesatellite-derived SSTs. Therefore,we adjusted the satellite SSTs forDampier based on the high degreeof covariation observed with an insitu SST record spanning 1983 to1986 where T in situ 1.67 T satellite18.67 ( r 0.95). Strong wind-in-duced upwelling confined to thesouthern coast of Java produces

cooler SSTs during the austral win-ter that are not recorded by the sat-ellite-derived SSTs, due to the 1spatial smoothingof these data. Weverify the anomalously cool SSTsrecorded by the coral in 1991 and1994usingmonthly averaged SSTs( ) derived from new along-track scanning radiometer (ATSR) dataaveraged over 0.5 latitude-longi-tude grids ( 37 ). Bars indicate peri-ods when upwelling (A) and freshwater runoff from mainland rivers (B) could potentially alter seawaterSr/Ca and reconstructed SSTs. The dashed line in (B) spans the drought of 1992 to 1994 used toestablish the correlation between coral Sr/Ca and 18 O (Fig. 2B).

Fig. 2. ( A ) Linear relations between coral Sr/Caand temperature for P. lutea colonies living in en-vironments that are suboptimal for coral growth.Regression lines are calculated with data pointsdefining the temperature maxima and minima of each data set, asdefined in Fig. 1. Sr/Ca ratiosforthe Java coral corresponding to SSTs 26C ( )are regressed against temperatures derived by

ATSR ( 37 ). The [Sr/Ca]-temperature ( T ) regressionlines are 10 3 [Sr/Ca] Java 10.78 0.0660 T ( r 0.80); 10 3 [Sr/Ca] Orpheus 10.73 0.0639 T ( r 0.98);and 10 3 [Sr/Ca] Dampier 10.68 0.0616 T ( r

0.99). (B ) Comparison of the relation betweencoral Sr/Ca-derived SSTs and 18 O for two P. lutea colonies calcifying at different rates in thesame reef environment at Orpheus Island, centralGreat Barrier Reef. The regression lines are[ 18O]1.3g cm 2 year 1 0.002 0.174 T ( r 0.98)and [ 18 O]2.5g cm 2 year 1 0.447 0.189 T ( r 0.98), where Sr/Ca-temperature is defined by themean of the three coral Sr/Ca-temperaturerelations.

R EPORTS

www.sciencemag.org SCIENCE VOL. 279 13 FEBRUARY 1998 1015


4/6

calibrated coral Sr/Ca temperatures and the18 O values can be confirmed during the

drought years of 1992 to 1994, which weremarked by stable seawater salinities, andpresumably stable seawater 18 O values(Fig. 1). The correlation between the coralSr/Ca and 18 O during these years is excel -lent; the average difference is 0.3C be -tween the two regression equations andtheir mean, despite the twofold differencein calcification rate (Fig. 2). The mean of the two 18 O-SST relations derived bycomparison with the coral Sr/Ca -SSTsagrees with an earlier 18 O-SST relationderived for P. lutea and P. lobata (23). Theresult indicates that consistent Sr/Ca -SSTand 18 O-SST relations can be derived formassive Porites colonies that are largely in -dependent of calcification rate.

We applied these calibration equations toskeletal Sr/Ca and 18 O measurements fora fossil Porites colony growing in the cen -tral Great Barrier Reef, Australia, to de -

termine the SST and oxygen isotopiccomposition of seawater soon after the endof the last deglaciation. Measurementswere made on the basal section of a core1.5 m long, 76 mm in diameter drilledthrough the center of a Porites micro-atoll6 m in diameter. The core spans 99 yearsof continuous coral growth and has a meanradiocarbon age ( 24) of 5350 190 yrB.P. (2 ). We compared the fossil coraldata directly with the calibration data for

the two modern Porites specimens growingin the same open -water reef environmenton the windward side of Orpheus Island.Several precautions were taken to ensurethe modern and fossil coral data are com -parable ( 25). Because of the long resi-dence times of Sr and Ca in ocean water(12), errors in the paleotemperature re -constructions for well -preserved mid -Ho -locene corals should be about the same asthose associated with comparisons amongmodern corals.

The paleotemperature indicated by thefossil coral Sr/Ca record shows that the meanSST 5350 years ago was 27.0C, which is1.2C warmer than the mean SST for theearly 1990s (Fig. 3). Terrestrial pollen andtree-line elevation records elsewhere in thetropical southwest Pacific indicate that theclimate was generally warmer from 7000 to4000 yr B.P. (26), consistent with our find -ing. The difference between the Sr/Ca and

18 O curves (Fig. 3) represents the residual18

O signal (18

O) ( 27) that can be used todefine the oxygen isotopic composition of seawater. The mean oxygen isotopic compo -sition of ambient seawater can be deter -mined with the 18 O residuals for the australwinters when transient changes in seawater

18 O are negligible ( 28) and central GreatBarrier Reef waters are in equilibrium withthe regional ocean. The line of best fitthrough the winter 18 O residuals in thefossil coral record defines an enrichment in

seawater 18 O of 0.47 0.13 per mil, relativeto modern values. If the mean oxygen isoto -pic composition of seawater at the end of thedeglaciation was the same as today, thereshould be no offset of the 18 O and Sr/Cacurves ( 18 O 0 per mil) for the fossilcoral. The 18 O enrichment defined by thewinters is not a result of a local decrease inprecipitation because the seasonal changesin the 18 O residuals for the fossil coral showthat, like today, runoff from mainland riverscarrying monsoonal rainfall with low 18 O/16 O ratios is restricted to the summer andearly autumn (December to May). Further -more, it is unlikely that changes in oceancirculation could have caused the 18 O en -richment. Subthermocline water masses inthe tropics have low 18 O/ 16 O ratios relativeto surface waters (29), so the imposition of an upwelling oceanic regime in the south -western Pacific would have the opposite ef -fect. Today, the only place in the Pacificwhere surface seawater 18 O approaches val -

ues 0.5 per mil higher than in the GreatBarrier Reef is beneath the southeast Pacificsubtropical high -pressure cell, where evapo -ration greatly exceeds precipitation ( 29).Advection of this 18 O-enriched water to thesouthwestern Pacific cannot be dismissed,but such a change would require a substan -tial reorganization of the regional oceanic -atmospheric circulation.

We propose that the 18 O enrichment isdriven by enhanced surface -ocean evapora -tion in the southwestern Pacific in responseto the higher SSTs at the end of the degla -ciation. Today, about 11 sverdrup (1 sver -drup 106 m3 s 1 ) of water evaporates fromthe warm surface of the tropics and subtrop -ics, of which about 1.5 sverdrup is transport -ed by the atmosphere and precipitated pole -ward of 40 (30). The export to higherlatitudes of this small portion of the evapo -rated water leads to an increase of 0.8 per milin the 18 O/ 16 O ratio of surface water of thetropical southwestern Pacific relative tomean ocean water ( 29). The 1C rise in SSTreported here would increase surface -oceanevaporation ( 31), particularly before the on -set of the monsoon when SSTs warm beyond27C to reach average summer maxima of 29.5C. Enhanced evaporation is evident in

the fossil coral record where the 18

O resid-

uals become progressively more positive inspring and reach maxima in November toDecember, immediately before the onset of monsoonal rainfall (Fig. 3). The observed18 O enrichment of surface seawater is con -sistent with general circulation modelsshowing that the poleward transport of watervapor can increase as the atmosphere warms(32). Theoretical and general circulationmodel studies also indicate that even subtlechanges in the distribution of tropical heat -ing, particularly warming off the equator,

Fig. 3. Comparison between calculated coral Sr/Ca (solid curves) and 18 O temperatures (upper curveswith circles) for the modern ( A ) and 5350 yr B.P. ( B ) Porites from Orpheus Island, central Great BarrierReef. Differences in seawater 18 O (lowercurves with circles),relativeto themodern mean, areobtainedby removal of the temperature component of the 18 O signal ( 18 O) ( 27 ). The horizontal lines show themean 18 O value of seawater foreach time-slice, as defined by theseven 18 O values (squares) fallingin the austral winters (vertical lines). 18 O values more negative than the mean define freshening of theocean surface by 18 O-depletedmonsoonal rainfall in summer, whereas 18 O valuesmore positive thanthe mean reflect the evaporative 18 O-enrichmentof thesurface ocean in spring. Theshaded areas showthe uncertainty in estimates of mean 18 O ( 0.13 per mil, 2 ) on the basis of differences between thedata points in Fig. 2B ( n 108) and the mean of the two 18 Otemperature regression lines. Theweighted mean regression line used to calculatethe 18 O temperaturesis [ 18 O]mean 0.146 0.179 T ( r 0.98).



5/6

can profoundly alter the intensity of theHadley circulation and the transport of wa -ter vapor and heat into the extratropics ( 33).Although the spatial distribution of thewarmer mid-Holocene SSTs needs to be es -tablished, it is possible that the equable cli -mates in the extratropical latitudes at thistime ( 26) are linked, in part, to an increasedlatent heat flux from the tropics.

The 18 O enrichment of surface seawateridentified here may help to reconcile thedifferent temperature histories for thedeglacial tropical Pacific given by coralSr/Ca thermometry and foraminiferal

18 O records. Taken together with the5C cooling indicated by Sr/Ca paleotem -peratures for late -glacial corals from thesouthwest Pacific ( 6), our results indicatethat the full amplitude of the glacial -Holocene temperature change may havebeen about 6C. A glacial -Holocene shiftin foraminiferal 18 O of 2.7 per mil isrequired to accommodate the 1.4 per mil

change attributed to the 6C deglacialwarming ( 34) and the 1.3 per mil for thegenerally accepted meltwater contribution(11). However, a recent compilation of the glacial -Holocene shift in foraminiferal

18 O values for 40 western Pacific sedi-ment cores in the latitude range 20S to20N ( 35) reveals an average shift of only1.6 0.7 per mil (2 ). The maximum

18 O shift given by the planktonic fora -minifer Globigerinoides sacculifer in high -quality cores from the tropical southwest -ern Pacific, near the Great Barrier Reef, is1.9 per mil (35), which is still 0.8 per milsmaller than the expected shift of 2.7 permil. The important point is that the 0.5per mil enrichment in surface water 18 Orevealed by the mid -Holocene coral recordwould reduce the expected glacial -Holo -cene shift in foraminiferal 18 O to 2.2 permil. Complete reconciliation of the fora -miniferal 18 O and coral Sr/Ca records isachieved if the meltwater contribution isreduced by 0.3 per mil, as indicated byrecent 18 O measurements of pore fluidsin deep -sea sediments ( 36).

The final 0.5 per mil change in seawater18 O during the late Holocene may not be

manifest in the 18 O record provided by

core-top foraminifera, considering that theuppermost late -Holocene sediment is com -

monly lost during the extraction of deep -seacores. Even in cores with perfect sedimentrecovery, more than one -half of the final 0.5per mil adjustment in seawater 18 O wouldbe masked in the foraminiferal 18 O recordsby the 1.2C cooling required to bring themid-Holocene SSTs to modern values.

The reconciliation of the foraminiferal18 O and coral Sr/Ca records proposed here

suggests that the constraint placed on trop -ical SSTs by foraminiferal 18 O should be

reassessed, at least for the tropical westernPacific. Changes in tropical SSTs, and therate of delivery of water vapor to higherlatitudes, could produce large and abruptchanges in global climate ( 9) during degla-ciation. Our results provide the basis fordetermining the response of the surface -ocean hydrologic balance to such changes,even in tropical oceanic settings whereboth temperature and seawater 18 Ochange in unexpected ways.

REFERENCES AND NOTES___________________________

1. J. Chappell and N. J. Shackleton, Nature 324 , 137(1986); N. J. Shackleton, Quat. Sci. Rev. 6 , 183(1987); J. Chappell et al. , Earth Planet. Sci. Lett.141 , 227 (1996).

2. Theoxygen isotopiccomposition of foraminiferal cal-cite is expressed in per mil where 18 O [(18 O/ 16 O)sample /(18 O/ 16 O)standard ] 1.

3. N. J. Shackleton, Nature 215 , 15 (1967); W. S.Broecker, Quat. Res. 26 , 121(1986);G. E.Birchfield,Paleoceanography 2 , 431 (1987).

4. CLIMAP Project Members, Geol. Soc. Am. MapChart Ser. MC-36 (1981).

5. T. P. Guilderson, R. G. Fairbanks, J. L. Rubenstone,Science 263 , 663 (1994).

6. J. W. Beck, J. Recy, F. Taylor, R. L. Edwards, G.Cabioch, Nature 385 , 705 (1996).

7. D. Rind and D. Peteet, Quat. Res. 24 , 1 (1985); M.Stute et al ., Science 269 , 379 (1995); L. G. Thomp-son et al ., ibid., p. 46.

8. W. S. Broecker, Paleoceanography 4 , 207 (1989);F.L. Norton, E. D.Hausman,M. B.McElroy, ibid. 12 ,15 (1997).

9. W. Broecker, Science 272 , 1902 (1996).10. S. deVilliers, B. K. Nelson, A. R. Chivas, ibid. 269 ,

1247 (1995).11. R. G. Fairbanks, Nature 342 , 637 (1989).12. J. W. Beck et al ., Science 257 , 644 (1992); Sr/Ca

ratios were determined by isotope dilution on a Finni-gan MAT 261 thermal ionization mass spectrometerwith 42 43 Ca- 84 Sr tracers. In our laboratory, the42 43 Ca double spike andthe natural 86 Sr/ 88 Sr ratioswere used to correct for mass fractionation. The re-producibility for coral Sr/Ca is 0.22% ( n 50, 2 ).

13. S. deVilliers, G. T. Shen, B. K. Nelson, Geochim.Cosmochim. Acta 58 , 197 (1994).

14. Slabs from corals were cut parallel to the axis of maximum growth and reduced to 2-mm thickness,ultrasonically cleaned, and dried at 40C. Samples150 to 350 m thick were then precisely shavedalong a continuous strip of 2 mm by 2 mm in cross-sectional area. Our microsampling procedure wasdesigned to provide a minimum of 50 samples perannual growth increment, despite large differencesin the annual extension rate among coral colonies,to ensure that the full range of the annual SST cyclewas well defined. Every second sample was ana-lyzed (about 25 samples per year), and these datawere routinely filtered with a two-point runningmean. For comparison with the coral records,blended ship- and satellite-derived SSTs ( 15 ) wereplotted at the same frequency (26 weekly incre-ments per year) and filtered with a two-point run-ning mean to simulate the coral-derived SSTs. Theinstrumental SSTs have been regressed to repro-duce insituSST measurements made by shipsandbuoys ( 15 ). Buoy measurements of SST are typi-cally made at a depth of 0.5 m, whereas ship mea-surements may be made as deep as 5 m. There-fore, the coral specimens were all collected from 2to 3 m below mean low sea level so that theycompare well with the satellite retrievals of SST.

15. R. W. Reynolds and T. M. Smith, J. Climate 7 , 929(1994).

16. M. K. Gagan, A. R. Chivas, P. J. Isdale, Geology 24 ,1009 (1996).

17. S. V. Smith, R. W. Buddemeier, R. C. Redalje, J. E.Houck, Science 204 , 404 (1979).

18. C.-C. Shen et al ., Geochim. Cosmochim. Acta 60 ,3849 (1996).

19. C. Alibert and M. T. McCulloch, Paleoceanography 12 , 345 (1997).

20. Oxygen-isotope analyseswere carriedout on the samesamplesby an automated individual-carbonate reaction(Kiel) device coupled with a Finnigan MAT 251 massspectrometer. Calibration to the Pee Dee belemnite(PDB)is throughNationalBureauof StandardsNBS-19. The reproducibility for coral 18 O is 0.06 per mil ( n 34, 2 ) for a typical 150- g sample.

21. Themeanannual extensionrate of each coral colonywas determined for at least 5 years of growth asdefined by the distance between annual densitybands revealed through x-radiography of coralslabs. Mean annual calcification rates were deter-mined by weighing a knownvolume ofcoralskeletonspanning 3 to 5 years of growth and calculating theaverage coral density per unit time. The average dif-ference between replicate coral density determina-tions was 0.02 g cm 3 .

22. J. M.Loughand D.J. Barnes, J. Exp. Mar. Biol. Ecol.211 , 29 (1997).

23. M. K. Gagan, A. R. Chivas, P. J. Isdale, Earth Planet.Sci. Lett. 121 , 549 (1994); M. K. Gagan and A. R.Chivas, Geophys. Res. Lett . 22 , 1069 (1995).

24. The conventional radiocarbon age of the sample(ANU-9923) is 5800 60 yr B.P. [M. Stuiver andH. A. Polach, Radiocarbon 19 , 353 (1977)]. The re-ported radiocarbon age wascorrectedfor theappar-ent age of local seawater by subtracting 450 35years from the conventional age [R. Gillespie andH. A. Polach, in Radiocarbon Dating , R. Berger andH. E. Suess, Eds.(Univ. of California Press, Berkeley,CA, 1979), pp. 404 421].

25. Special care was taken to obtain comparable datasets by matchingclosely the growth environments of the modern and fossil corals. For example, the pa-leo-depth of the fossil coral is well constrained be-cause the tops of open-water micro-atolls grow toabout themeanlow-tidelevel. Thefossil coral recordwas extracted from the basal section of the coralcore 1.4 m below mean paleo-low tide, whereasthe tops of the modern corals were 2 m belowmean low tide. The coral growth rates are also com-parable; the mean annual extension rate (15 mmyear 1 ) and calcification rate (1.6 g cm 2 year 1 ) forthe fossil coral is within the range used to calibratethe modern corals.

26. COHMAP Members, Science 241 , 1043 (1988); M. S.McGlone, A. P. Kershaw, V. Markgraf, in El Nin o: His-torical andPaleoclimatic Aspectsof the Southern Oscil- lation, H. F. Diaz and V. Markgraf, Eds. (CambridgeUniv. Press, New York, 1994), pp. 435462.

27. 18 O 18 O/ T [T 18O T Sr/Ca ], where 18 O/ Tis the empirically derived temperature-dependentoxygen isotope fractionation ( 0.18 per mil per de-gree Celsius) forthe Porites investigated in this studyand elsewhere ( 23 ).

28. M. T. McCulloch, M. K. Gagan, G. E. Mortimer, A. R.Chivas, P. J. Isdale, Geochim.Cosmochim. Acta 58 ,2747 (1994).

29. H. Craig and L. Gordon, in Stable Isotopes in Oce- anic Studies and Paleotemperatures , E. Tongiorgi,Ed. (Spoleto, Pisa, Italy, 1965), pp. 9130; H. G.Ostlund, H. Craig, W. S. Broecker, D. Spencer,GEOSECS Atlantic, Pacific and Indian Ocean Expe-ditions. Shorebased Data and Graphics 7 (NationalScience Foundation, Washington, DC, 1987).

30. W. S. Broecker and G. H. Denton, Geochim. Cos- mochim. Acta 53 , 2465 (1989); F. L. Norton, E. D.Hausman, M. B. McElroy, Paleoceanography 12 , 15(1997).

31. S. Bony, J.-P. Duvel, H. L. Treut, Climate Dyn. 11 ,307 (1995).

32. D. Rind, J. Atmos. Sci. 44 , 3235 (1987); D. Rind, R.Goldberg, J. Hansen, C. Rosenzweig, R. Ruedy, J.Geophys. Res. 95 , 9983 (1990).

33. A. Y. Hou, J. Atmos. Sci. 50 , 3553 (1993); R. S.Lindzen and W. Pan, Climate Dyn. 10 , 49 (1994).

34. The oxygen-isotope thermometer for calcite pre-cipitation at tropical temperatures in the planktonicforaminifer G. sacculifer is 0.23 per mil per de-gree Celsius [J. Erez and B. Luz, Geochim. Cos- mochim. Acta 47 , 1025 (1983)].

R EPORTS

www.sciencemag.org SCIENCE VOL. 279 13 FEBRUARY 1998 1017


6/6

35. J. I. Martinez, Palaeogeogr. Palaeoclimatol. Palaeo-ecol. 112 , 19 (1994); , P. DeDeckker, A. R.Chivas, Mar. Micropaleontol. 32 , 311 (1997).

36. D. P. Schrag, G. Hampt, D. W. Murray, Science272 , 1930 (1996).

37. The along-track scanning radiometer (ATSR) dataare supplied by the European Space Agency andhavebeen analyzedby the Rutherford AppletonLab-oratory, UK.

38. We thank I. Ward for assistance with coral samplepreparation and mass spectrometry; S. Anker, S. Vellacott, P. Walbran, and W. Hantoro for help withcollecting corals; and W. Beck, D. Schrag, B. Lins-ley, D. Battisti, and K. Trenberth for valuable discus-sions. Financialsupport was provided by the Austra-lian National Greenhouse Advisory Committee.

3 October 1997; accepted 23 December 1997

Simulated Increase of Hurricane Intensitiesin a CO 2 -Warmed Climate

Thomas R. Knutson,* Robert E. Tuleya, Yoshio Kurihara

Hurricanes can inflict catastrophic property damage and loss of human life. Thus, it isimportant to determine how the character of these powerful storms could change inresponse to greenhouse gasinduced global warming. The impact of climate warmingon hurricane intensities was investigated with a regional, high-resolution, hurricaneprediction model. In a case study, 51 western Pacific storm cases under present-dayclimate conditions were compared with 51 storm cases under high-CO 2 conditions.Moreidealized experiments were also performed. The large-scale initial conditions were de-rived from a globalclimate model.For a seasurface temperature warming of about 2.2C,

the simulations yielded hurricanes that were more intense by 3 to 7 meters per second(5 to 12 percent) for wind speed and 7 to 20 millibars for central surface pressure.

G reenhouse gasinduced climate warmingcould affect hurricanes in a number of ways,including changing their intensity ( 1, 2),frequency (35), and locations of occur -rence. Given the potential for catastrophicdamage and loss of life from these storms,any such changes could have important so -cietal consequences. In this study, we ex -amine only the question of possible changesin storm intensity due to climate warming.

Theoretical models of hurricane intensi -ty predict that the maximum potential in -tensity (MPI) of hurricanes will increase ina warmer climate ( 1, 2), although thesetechniques, which are based on thermody -namical considerations, contain many as -sumptions and caveats ( 2, 6, 7). Globalclimate models attempt to simulate the cli -mate, including tropical stormlike fea -tures, by integrating dynamical and thermo -dynamical equations in three dimensions.To date, global models have provided sug -gestive, but not highly convincing, indica -tions of increased hurricane intensities in a

warmer climate ( 3, 4). However, the coarseresolution of these global models precludestheir simulation of realistic hurricane struc -ture. A 1995 assessment by the Intergov -ernmental Panel on Climate Change ( 8)concludes that . . . it is not possible to saywhether the . . . maximum intensity of tropical cyclones will change because of

increased greenhouse gas concentrations. Inthe present study, the relation between hur -ricane intensity and climate change wasexplored with a regional, high -resolution,hurricane prediction model. We focused onthe northwest tropical Pacific region, wherethe strongest typhoons (the term used inthe northwestern Pacific for hurricanes) areobserved in the present climate.

In our case study approach, we selected51 tropical storm cases from a control cli -mate simulation of a global climate modeland 51 cases from a high-CO 2 climate sim-ulation ( 9). The global model used was theGeophysical Fluid Dynamics Laboratory(GFDL) R30 coupled ocean -atmosphereclimate model ( 1012), which has resolu -tion of about 2.25 latitude by 3.75 longi -tude. For the high -CO 2 cases, we selectedstorms from years 70 to 120 of a 1%-per-year CO 2 transient experiment, correspond -ing to CO 2 increases ranging from a factorof 2.0 to 3.3. Tropical stormlike features(weaker and much broader than in real -

world storms) have previously been ana-

lyzed in an R30 global atmospheric modelvery similar to that used here ( 5, 13). Theselected storm cases were then rerun as5-day forecast experiments with the use of the high -resolution GFDL Hurricane Pre -diction System ( 14), which is currently usedat the U.S. National Centers for Environ -mental Prediction (NCEP). This model hasa maximum resolution in the storm regionof 1/6 or about 18 km ( 15). Before begin-ning each hurricane model simulation, thecrudely resolved global model storm (but

not the background environment) was fil -tered from the global model fields and re -placed by a more realistic initial vortex ( 16,17). This initial vortex replacement proce -dure is analogous to that used for operation -al hurricane prediction at NCEP. The stormintensity distributions of the control andhigh -CO 2 case studies were then compared.Sea surface temperatures (SSTs) were heldfixed during the hurricane model experi -ments. The SSTs and initial environmental

Geophysical FluidDynamics Laboratory/NationalOcean-ic and Atmospheric Administration, Post Office Box 308,Princeton, NJ 08542, USA.

*To whom correspondence should be addressed. E-mail:[email protected]

A

B

C

Fig. 1. Geographical distribution of the maximumsurface wind speeds (in meters per second) ob-

served during 19711992 ( A ) and simulated ( Band C ) for tropical storms in the northwest Pacificbasin. Observations are from the Joint TyphoonWarning Center(Guam) as compiled by C. J. Neu-mann in 1993, available from the National Centerfor Atmospheric Research at www.scd.ucar.edu/ dss (ds824.1). The simulated distributions arebased on 71 case studies each under control (B)and high-CO 2 (C) conditions; results from 20 pre-liminary cases under each condition ( 9 ) were in-cluded in order to increase spatial coverage.Blank (white) regions denoteareas where no trop-ical storms were reported during 19711992 (A)or none occurred in the case studies [(B) and (C)].


Holoceno Medio Corales Australia

Documents

Transcript of Holoceno Medio Corales Australia